Unverified50% confidenceFactExact time
Around 2019, multiple research teams used statistical language models and neural networks to analyze the distribution of Voynich Manuscript symbols; some concluded that the underlying language may exhibit statistical properties of natural language, but no decipherment result has gained broad academic recognition to this day.
1
Sources
50%
Confidence
Long-term
Relevance
9/22/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedAnthropic's research team successfully identified feature vectors inside large language models corresponding to concepts such as deception and sycophancy using sparse autoencoders in 2023–2024.74% similarUnverified研究表明模型在某语言上的文化理解深度与该语言的训练数据量级之间存在近似对数线性关系73% similarUnverifiedAbliteration技术源自2024年开源社区对大语言模型内部表征的研究，理论基础是线性表征假说72% similarUnverifiedTraditional n-gram models and neural language models trained on a single language assign extremely low probability scores to cross-lingual word sequences, causing the decoder to tend to forcibly map foreign-language words to phonetically similar words in the target language.71% similarVerifiedThe emergent ability phenomenon was systematically documented by DeepMind researchers in the 2022 paper 'Emergent Abilities of Large Language Models'.71% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/940724API
curl https://kongchang.com/api/v1/knowledge/claims/940724MCP
get_claim(id=940724)